MD1000 connected with two SAS cables in unified mode

Michael_Christensen2 at Dell.com Michael_Christensen2 at Dell.com
Tue Sep 16 01:28:33 CDT 2008


   2. Re: MD1000 connected with two SAS cables in unified mode
      (Jeff Ewing)
  
Message: 2
Date: Mon, 15 Sep 2008 14:01:16 +1000
From: Jeff Ewing <jewing at aconex.com>
Subject: Re: MD1000 connected with two SAS cables in unified mode
To: linux-poweredge at lists.us.dell.com
Message-ID: <20080915040116.GD5462 at mail.aconex.com>
Content-Type: text/plain; charset=us-ascii

>From what I understand of the MD1000 EMMs, there is failover, but only
for the management communications to the chassis, there is no data path
failover.. so losing the active EMM/connection will result in the PERC
needing to re-init (host reboot). This limitation is determined by the
PERC..
> 
> Maybe anyone else who's done this before can comment on it's success..
> 

I had a spare MD1000 so I performed some data path tests.

* Both cables connected in unified mode. 

  The primary path was disconnected.

  - The data was inaccessible.

* Rebooted 

  - The data was accessible down the alternate path.

* Shutdown, reconnected the primary cable and restarted. 

   -The data was accessible (although the RAID array went into 
    a rebuilding state)


For me this appears to provide the required path redundancy (using
a reboot). A reboot is an acceptable cost compared with organizing
onsite technicians at the Data Centre in the case of a path failure.

Jeff


Hi Jeff,

I strongly suggest sticking to the supported cabling scenarios, as I
have had quite a few people contact me with hosed data after having used
the same configuration as the one you proposing is okay.

EMM Cabling Diagram in Unified and Split Mode
http://support.dell.com/support/edocs/systems/md1000/en/HOM/operate.htm#
wp1057117


//Michael




More information about the Linux-PowerEdge mailing list